Comparison of performance of four instruments in evaluating the effects of salmeterol on asthma quality of life

Quality of life measures are increasingly used as important efficacy endpoints in studies of drugs for asthma. The purpose of this study was to assess both the sensitivity to change and the construct validity of four different quality of life instruments in patients with asthma. In a double-blind, parallel group study, 120 moderate asthma patients, aged between 18-70 yrs, received either inhaled salmeterol 50 micrograms b.i.d. or inhaled salbutamol 400 micrograms b.i.d. In addition to respiratory outcomes, quality of life was measured at a 6 weeks follow-up using: 1) Asthma Quality of Life Questionnaire (AQLQ); 2) Living With Asthma Questionnaire (LWAQ); 3) Sickness Impact Profile (SIP); 4) Rating Scale (RS); and Standard Gamble (SG) utilities. Salmeterol led to significant improvements over salbutamol on virtually all clinical outcomes. Although all the quality of life instruments showed the same trend in favour of salmeterol, only the AQLQ and RS utilities showed significantly greater improvement on salmeterol than on salbutamol. Except for the AQLQ, the correlation between change in lung function and change in quality of life was generally low. Whereas, the AQLQ correlated well with the patient's overall assessment of efficacy (r = 0.64), the LWAQ, SIP and utilities failed to show such a correlation. The AQLQ showed the best correlation with symptom scores. The cross-sectional correlation between the AQLQ and the LWAQ was 0.73, whereas the longitudinal correlation was only 0.29. The SG generally showed poor correlation with other measures, including the RS. In conclusion, patients given salmeterol showed a greater improvement in quality of life compared to patients given salbutamol. Of the disease-specific questionnaires the Asthma Quality of Life Questionnaire was found to be more responsive to change than the Living With Asthma Questionnaire and showed greater validity. Of the generic instruments, the rating scale utilities were most responsive. The Standard Gamble showed poor correlation with other measures.